Candles and Market Hours

Veyronn does not synthesize market movement.

Historical backfill

Pyth Benchmarks supplies authentic OHLC history for the market's configured benchmark symbol. The BFF normalizes these records and Railway Postgres persists them, so the product builds its own durable history rather than depending on every client repeatedly backfilling a third-party endpoint.

Live candle construction

The keeper's Hermes stream persists accepted observations. For each interval bucket:

• the first observation becomes open
• the maximum becomes high
• the minimum becomes low
• the latest becomes close
• observation count and timestamps support continuity checks

The frontend combines persisted history with the newest live observation to extend the active candle.

Interval aggregation

Stored observations and base candles can be aggregated into 1m, 5m, 15m, 1h, 4h, and 1d views. RSI and moving averages are derived from authentic closes.

Closed markets

Real-world asset feeds can stop or report a closed status outside their underlying session. Veyronn:

• keeps historical candles visible
• labels the market closed or stale
• blocks new risk
• avoids inserting flat or fabricated candles
• resumes live candle construction when acceptable observations return

Pyth authentication transition

Pyth has announced authentication requirements for public data access beginning August 18, 2026. Veyronn supports server-only API keys and persists its own candle history so adding credentials does not change the frontend data contract. Keys belong on Railway, never in NEXT_PUBLIC_* variables.